Temperature Drift Compensation of Fiber Optic Gyroscopes Based on an Improved Method

Abstract

This study proposes an improved multi-scale permutation entropy complete ensemble empirical mode decomposition with adaptive noise (MPE-CEEMDAN) method based on adaptive Kalman filter (AKF) and grey wolf optimizer-least squares support vector machine (GWO-LSSVM). By establishing a temperature compensation model, the gyro temperature output signal is optimized and reconstructed, and a gyro output signal is obtained with better accuracy. Firstly, MPE-CEEMDAN is used to decompose the FOG output signal into several intrinsic mode functions (IMFs); then, the IMFs signal is divided into mixed noise, temperature drift, and other noise according to different frequencies. Secondly, the AKF method is used to denoise the mixed noise. Thirdly, in order to denoise the temperature drift, the fiber gyroscope temperature compensation model is established based on GWO-LSSVM, and the signal without temperature drift is obtained. Finally, the processed mixed noise, the processed temperature drift, the processed other noise, and the signal-dominated IMFs are reconstructed to acquire the improved output signal. The experimental results show that, by using the improved method, the output of a fiber optic gyroscope (FOG) ranging from −30 °C to 60 °C decreases, and the temperature drift dramatically declines. The factor of quantization noise (Q) reduces from 6.1269 × 10−3 to 1.0132 × 10−4, the factor of bias instability (B) reduces from 1.53 × 10−2 to 1 × 10−3, and the factor of random walk of angular velocity (N) reduces from 7.8034 × 10−4 to 7.2110 × 10−6. The improved algorithm can be adopted to denoise the output signal of the FOG with higher accuracy.

Keywords

• fiber optic gyroscope (FOG)
• multi-scale permutation entropy complete ensemble empirical mode decomposition with adaptive noise (MPE-CEEMDAN)
• adaptive Kalman filter (AKF)
• grey wolf optimizer-least squares support vector machine (GWO-LSSVM)
• temperature compensation